Simple Shape Detection

Polygon detection and classification using OpenCV contour analysis.

How It Works

1. Convert input image to grayscale and apply binary thresholding
2. Find contours with cv2.findContours
3. Filter out small contours (noise) by area
4. Approximate each contour to a polygon using the Douglas-Peucker algorithm
5. Classify by vertex count and geometry:
   • 3 vertices → Triangle
   • 4 vertices → Square or Rectangle (distinguished by aspect ratio)
   • 5 vertices → Pentagon
   • 6 vertices → Hexagon
   • More → Circle
6. Draw each shape with a distinct color and label

Usage

pip install opencv-python numpy
python simple_shape_detection.py              # defaults to polygons.png
python simple_shape_detection.py myimage.png  # custom image

Dependencies

• OpenCV
• NumPy

License

MIT